H2O diffusion in rhyolitic melts and glasses

Citation
Yx. Zhang et H. Behrens, H2O diffusion in rhyolitic melts and glasses, CHEM GEOL, 169(1-2), 2000, pp. 243-262
Citations number
34
Categorie Soggetti
Earth Sciences
Journal title
CHEMICAL GEOLOGY
ISSN journal
00092541 → ACNP
Volume
169
Issue
1-2
Year of publication
2000
Pages
243 - 262
Database
ISI
SICI code
0009-2541(20000815)169:1-2<243:HDIRMA>2.0.ZU;2-J
Abstract
H2O diffusion plays a major role in bubble growth and volcanic eruption. We report a comprehensive study of H2O diffusion in rhyolitic melts and glass es. This new study and previous investigations together cover a wide range of conditions: 400-1200 degrees C, 0.1-810 MPa, and 0.1-7.7 wt.% total H2O content (H2Ot). In order to constrain how the diffusivity depends on H2Ot, both the diffusion-couple experiments and the dehydration experiments are c arried out in a cold-seal vessel (CSV), an internally heated pressure vesse l, and a piston cylinder. H2O concentration profiles are measured by infrar ed (IR) spectroscopy. Although there are still some experimental and analyt ical difficulties, our data represent a major improvement over earlier data . The diffusion data have been used to quantify H2O diffusivity as a functi on of temperature, pressure, and H2Ot. Assuming that molecular H2O (H2Om) i s the diffusing species, the H2Om diffusivity (in mu m(2)/s) can be express ed as: DH2Om = exp[14.08-13,128/T-2.796/T) + (-27.21 + 36,892/T + 57.23P/T)X], where T is in Kelvin, P is in mPa, and X is the mole fraction of H2Ot on a single oxygen basis. The pressure dependence is not so well-resolved compar ed to the dependence on T and X. The dependence of DH2Om on X increases wit h increasing pressure. The results are consistent with the data of Nowak an d Behrens (1997) [Nowak, M., Behrens, H., 1997. An experimental investigati on on diffusion of water in haplogranitic melts. Contrib. Mineral. petrol. 126, 365-376.], but different from the assumption of Zhang et al. (1991a) [ Zhang, Y., Stolper, E.M., Wasserburg, G.J., 1991a. Diffusion of water in rh yolitic glasses. Geochim. Cosmochim. Acta 55, 441-456.], because the depend ence cannot be resolved from their low-H2Ot diffusion data, and because the dependence is not so strong at low pressures. The activation energy for H2 Om diffusion decreases as H2Ot increases and depends on P (increases with P at X < 0.05 and decreases with P at X > 0.05). The results roughly reconci le the different activation energies of Zhang et al. (1991a) and Nowak and Behrens (1997). The total (or bulk) H2O diffusivity (DH2Ot) can be calculat ed from DH2Ot = DH2Ot dX(m)/dX, where X-m is the mole fraction of H2Om. Thi s approach can reproduce the DH2Ot values to within a factor of 2 in the ra nge of 400-1200 degrees C, 0.1-810 MPa, and 0-7.7% H2Ot. An explicit formul a for calculating D-H2Ot at H2Ot less than or equal to 2% is: D-H2Ot = C/C(0)exp(10.49 - 10,661/T - 1.772P/T), where C is H2Ot content by weight, and C-0 equals 1% H2Ot. A formula for ca lculating D-H2Ot at all conditions covered by this work is: DH2Ot = Xexp(m){1+ exp[56 + m + X{--34.1 + 44,620/T + 57.3P/T) - root X(0.0 91 + 4.77 x 10(6)/T-2)]}, where m = -20.79 - 5030/T- 1.4P/T. The diffusivities obtained in this work can be used to model bubble growth in explosive and nonexplosive rhyolitic volcanic eruptions in all commonly encountered T, P, and H2Ot conditions. ( C) 2000 Elsevier Science B.V. All rights reserved.